Artificial Intelligence 10 years from now

By Nuno Moniz and João Vinagre, researchers at the Artificial Intelligence and Decision Support Laboratory.

The birth of the concept of Artificial Intelligence concept is commonly connected to a workshop held at the University of Dartmouth, in the summer of 1956: the Dartmouth Summer Research Project on Artificial Intelligence. The event’s proposal, established in the previous year, stated that one of the primary objectives was to understand if machines could describe and simulate one or all aspects of human learning. Despite its subtle success, this event launched an endeavour with more than 70 years, and contributed decisively to the evolution of the concept into an area of research.

There have been many people predicting a great future for Artificial Intelligence since its dawn; like a future where machines would replace the manual and intellectual work of humans. Undoubtedly, more than a thought-provoking and hypothetical scenario, this is a crucial and current debate, with serious implications for work organisation, a key aspect to our interpersonal relationships and to society itself. However, and maybe because AI is frequently presented as a hypothetical scenario, one can often become dazzled, due to our (common) lack of formal skills in social and political analysis. The truth is that most people were not educated in that sense, so we are excused.

The most common variant of these propaganda-statements is predicting the possibilities of Artificial Intelligence in the future. The truth is that no one can guess the degree of AI maturity 10 years from now. Let alone, the capabilities of AI in 25 years, or after a generation. In general, whenever there is a glimpse of what Artificial Intelligence will do in the future, there are good reasons to become suspicious. Not for any personal reasons, but because all the previous ones were wrong. All predictions in this area have at least one problem in common: they assume that the evolution of things involving social relations, like human skills, scientific development and understanding, can be anticipated in the medium and long term. This is a common aspect in the predictions that Artificial Intelligence systems will replace physical and intellectual tasks. As a rule, they perceive work, and even society, as something sealed, lacking the capacity to adapt and evolve. If there is anything we can say for sure – even without specific training in social and political sciences – is that throughout history, the technological advances led to complex and unpredictable economic, social, and political changes.

In order to get things clear, the distrust and fear by many people concerning the impact of Artificial Intelligence on their lives today, as well as on the future, is clearly justified. It is true that solutions based on Artificial Intelligence have the potential to replace tasks that, until now, were exclusively performed by humans. However, it is also true that the path we are taking today is more focused on coexistence and collaboration between humans and intelligent systems. Nevertheless, the prospect of a future in which automation could have a considerable impact on the employment expectations of an entire generation – probably older and with low higher education rates – is something one should consider and prepare for. In fact, it is the only clear advantage of this exercise of predicting the future of Artificial Intelligence: it can give us enough time to think about it and prepare ourselves.